RESEARCH ARTICLE

Quantitative assessment of fecal

contamination in multiple environmental

sample types in urban communities in Dhaka,

Bangladesh using SaniPath microbial

approach

Nuhu AminID1*, Mahbubur Rahman1, Suraja Raj2, Shahjahan Ali1, Jamie Green2,

Shimul Das1, Solaiman Doza1, Momenul Haque Mondol1,3, Yuke Wang2, Mohammad

Aminul Islam4,5, Mahbub-Ul Alam1, Tarique Md. Nurul Huda1, Sabrina Haque6,

Leanne Unicomb1, George Joseph6, Christine L. Moe2

1 Infectious Disease Division, International Centre for Diarrhoeal Disease Research, Bangladesh (icddr,b),

Dhaka, Bangladesh, 2 Center for Global Safe Water, Sanitation, and Hygiene, Emory University, Atlanta,

Georgia, United States of America, 3 Department of Statistics, University of Barishal, Barishal, Bangladesh,

4 Laboratory Sciences and Services Division, International Centre for Diarrhoeal Disease Research,

Bangladesh (icddr,b), Dhaka, Bangladesh, 5 Paul G Allen School for Global Animal Health, Washington State

University, Pullman, Washington, United States of America, 6 Water Global Practice, The World Bank,

Washington DC, United States of America

* nuhu.amin@icddrb.org

Abstract

Rapid urbanization has led to a growing sanitation crisis in urban areas of Bangladesh and

potential exposure to fecal contamination in the urban environment due to inadequate sani-

tation and poor fecal sludge management. Limited data are available on environmental fecal

contamination associated with different exposure pathways in urban Dhaka. We conducted

a cross-sectional study to explore the magnitude of fecal contamination in the environment

in low-income, high-income, and transient/floating neighborhoods in urban Dhaka. Ten sam-

ples were collected from each of 10 environmental compartments in 10 different neighbor-

hoods (4 low-income, 4 high-income and 2 transient/floating neighborhoods). These 1,000

samples were analyzed with the IDEXX-Quanti-Tray technique to determine most-proba-

ble-number (MPN) of E. coli. Samples of open drains (6.91 log10 MPN/100 mL), surface

water (5.28 log10 MPN/100 mL), floodwater (4.60 log10 MPN/100 mL), produce (3.19 log10

MPN/serving), soil (2.29 log10 MPN/gram), and street food (1.79 log10 MPN/gram) had the

highest mean log10 E. coli contamination compared to other samples. The contamination

concentrations did not differ between low-income and high-income neighborhoods for

shared latrine swabs, open drains, municipal water, produce, and street foodsamples. E.

coli contamination levels were significantly higher (p <0.05) in low-income neighborhoods

compared to high-income for soil (0.91 log10 MPN/gram, 95% CI, 0.39, 1.43), bathing water

(0.98 log10 MPN/100 mL, 95% CI, 0.41, 1.54), non-municipal water (0.64 log10 MPN/100

mL, 95% CI, 0.24, 1.04), surface water (1.92 log10 MPN/100 mL, 95% CI, 1.44, 2.40), and

floodwater (0.48 log10 MPN/100 mL, 95% CI, 0.03, 0.92) samples. E. coli contamination

PLOS ONE | https://doi.org/10.1371/journal.pone.0221193 December 16, 2019 1 / 21

a1111111111

a1111111111

a1111111111

a1111111111

a1111111111

OPEN ACCESS

Citation: Amin N, Rahman M, Raj S, Ali S, Green J,

Das S, et al. (2019) Quantitative assessment of

fecal contamination in multiple environmental

sample types in urban communities in Dhaka,

Bangladesh using SaniPath microbial approach.

PLoS ONE 14(12): e0221193. https://doi.org/

10.1371/journal.pone.0221193

Editor: Vassilis G. Aschonitis, Hellenic Agricultural

Organization - Demeter, GREECE

Received: July 30, 2019

Accepted: November 19, 2019

Published: December 16, 2019

Copyright: © 2019 Amin et al. This is an open

access article distributed under the terms of the

Creative Commons Attribution License, which

permits unrestricted use, distribution, and

reproduction in any medium, provided the original

author and source are credited.

Data Availability Statement: All relevant data are

within the paper and its Supporting Information

files.

Funding: The study was financially supported by

the Bill & Melinda Gates Foundation (grant no.

00010161) through the Rollins School of Public

Health at Emory University. icddr,b acknowledges

with gratitude the commitment of the Bill &

Melinda Gates Foundation and Rollins School of

Public Health, Emory University to its research

were significantly higher (p<0.05) in low-income neighborhoods compared to transient/float-

ing neighborhoods for drain water, bathing water, non-municipal water and surface water.

Future studies should examine behavior that brings people into contact with the environ-

ment and assess the extent of exposure to fecal contamination in the environment through

multiple pathways and associated risks.

Introduction

Globally, an estimated 24% of the total disease burden and 23% of all deaths are attributed to

environmental factors [1]. Inadequate sanitation and unsafe fecal sludge management threaten

public health through fecal contamination in the environment in many low- and middle-

income countries [2,3]. Dhaka, the capital of Bangladesh, is one of the most densely populated

cities in the world [4]. Fecal contamination in the environment is common in Dhaka neigh-

borhoods due to many factors, including poor sanitation and sewerage systems, rapid

unplanned urbanization, frequent flooding [5,6], and inefficient solid waste management

[7,8]. Recent studies in urban Dhaka [9] and Khulna [10] also found that about 80% of fecal

sludge from on-site pit latrines is not safely managed [11].

Limited studies have been conducted to quantify levels of fecal contamination in different

environmental compartments in urban Dhaka [12–15]. Direct ingestion of fecal contamina-

tion through contaminated drinking water has been studied extensively both at household and

community levels in urban Bangladesh by measuring fecal indicator bacteria [16–19]. Other

exposure pathways in urban Bangladesh, including contaminated soil [13], market produce

[12], and street food [14] have been linked to adverse health outcomes such as diarrhea, envi-

ronmental enteric dysfunction, and stunting [20,21]. Yet the contribution of these pathways to

total fecal exposure remains understudied. Most urban studies have had small sample sizes,

studied few communities, and targeted only a limited number of specific environmental com-

partments (i.e., market produce, soil, or street food), which are unlikely to provide a complete

picture of the environmental fecal contamination levels in those communities. To inform evi-

dence-based decision-making processes, policymakers, local government administrators, and

local NGOs need data on the full range of fecal contamination pathways in order to more

effectively prioritize and target interventions.

The SaniPath Exposure Assessment Tool quantitatively assesses exposure to fecal contami-

nation via multiple pathways using a combination of microbiological data on environmental

samples and information on the frequency of behaviors involving exposure to each environ-

mental pathway [20,21]. We conducted a cross-sectional study to investigate the levels of envi-

ronmental fecal contamination in different environmental compartments in 10neighborhoods

in urban Dhaka using the SaniPath Exposure Assessment Tool [20]. In addition, information

on relevant physical and demographic characteristics of the study neighborhoodswas

collected.

Methods

Enrollment of study neighborhoods

Before study site selection, we conducted a stakeholder meeting and shared our protocol with

local collaborators, partners, policymakers, and national and international NGOs to develop

neighborhood selection criteria based on the water and sanitation context in urban Dhaka. We

SaniPath fecal exposure pathways urban Dhaka, Bangladesh

PLOS ONE | https://doi.org/10.1371/journal.pone.0221193 December 16, 2019 2 / 21

efforts. We acknowledge the World Bank

Bangladesh Country Office team for their efforts in

partnering, workshops, and contributing to

sampling/study design. icddr,b is also grateful to

the Governments of Bangladesh, Canada, Sweden

and the UK for providing core/unrestricted support.

The funders had no role in study design, data

collection and analysis, decision to publish, or

preparation of the manuscript.

Competing interests: The authors have declared

that no competing interests exist.

selected neighborhoods based on socio-economic status, stability of the population (i.e., per-

manent vs. floating/transient population), nature of the housing and WASH infrastructure

and services (i.e., unstructured vs. structured slums, and non-slums with poor WASH facili-

ties/services, and non-slums with improved WASH facilities/services) and varied geographic

locations [Dhaka South City Corporation (DSCC) and Dhaka North City Corporation

(DNCC)] (S1 Fig).

In 2011, the Dhaka City Corporation (DCC) was divided and re-created as DSCC and

DNCC under an amendment act [22]. In this study, we collectedan equal number of samples

from neighborhoods in each corporation to explore differences in E. coli concentrations

between the city corporations. We selected 10 neighborhoods from urban Dhaka (five from

each city corporation) between April and June 2017: four low-income neighborhoods, two

“floating” communities with transient populations, and four middle- to high-income neigh-

borhoods (S1 Table). We used the Bangladesh Bureau of Statistics (BBS) 2014 [23] slum list to

select the low-income neighborhoods and floating communities for this assessment. We

enrolled low-income neighborhoods that included at least 300 household compounds each

and categorized them into unstructured” (Badda and Hazaribagh) and “structured” (Kalshi

and Shampur) slums. “Unstructured” slums (poorest wealth quintile, average monthly house-

hold income: UD$ 104 [1US$ = 85 BDT]) had poorly structured housing (woods, bamboo, tin

etc.), poor water distribution systems (i.e., through flexible pipes) and poor sanitation facilities

(i.e., mostly hanging toilets) compared to structured slums. “Structured” slums (second lowest

wealth quintile, average monthly household income: UD$ 145) had permanent household

structures, >20 hours water supply per day, and shared latrine facilities. We selected the Gab-

toli bus terminal and Kamalapur railway station areas as floating communities (poorest quin-

tile: landless people who do not have any land for cultivation or homestead) because of the

transient populations who live in these areas and do not have permanent dwellings [24]. Four

middle- and high-income communities were selected from two separate elite communities

(Gulshan and Dhanmondi [highest wealth quintile, average monthly household income: UD$

345]), one commercial/business area (Motijhil [middle and forth wealth quintile, average

monthly household income: UD$ 146 and 225 respectively]), and one newly developed neigh-

borhood (Uttarkhan [high and middle wealth quintiles]) from urban Dhaka (S1 Fig and S1

Table) [25].

In each neighborhood, we conducted one key informant interview (KII) with either a city

official (i.e. city corporation staff and ward commissioners) or a community leader (i.e., local

political leaders, religious leaders, or NGO workers/representatives) who had lived or worked

in the selected neighborhood for more than five years and had a good understanding about the

water, sanitation, and hygiene (WASH) facilities and practices of the neighborhood.

A total of 1,000 environmental samples (10 neighborhoods x 10 sample types x 10 samples

per type) were collected. The sample types included: 1) swabs from the walls (100 cm2 on flat

surfaces) and door handles of shared/communal and/or public latrines accessed by any neigh-

borhood residents, 2) soil/sand/mud frompublic areas where people gather and children com-

monly play, 3) open drain water from an open channel, carrying liquid and solid waste,

including rainwater, floodwater, and wastewater from toilets and household activities, from

locations where community people and children commonly come into contact, 4) water from

both municipal and non-municipal water supplies that was used to bathe children (i.e., stored

water from the municipal water supply, shallow tubewell water, or surface water), 5) municipal

drinking water distributed through the Saidabad surface water treatment plant (with or with-

out additional water treatment at the neighborhood level by booster chlorination [ie. in-line

chlorine injectors]) and from deep bore wells attached with or without booster chlorinator

[26], (both legal and illegal connections) supplied by Dhaka Water Supply and Sewerage

SaniPath fecal exposure pathways urban Dhaka, Bangladesh

PLOS ONE | https://doi.org/10.1371/journal.pone.0221193 December 16, 2019 3 / 21

Authority (WASA) and accessed through piped water into compounds (including flexible

pipes) and public taps/stand posts that are provided by the government or managed by some-

one in the community, 6) non-municipal drinking water (20 L commercially available jars

from the local vendors or submersible pumps connected to a deep borehole), 7) surface water

from community ponds and/or lakes, 8) floodwater that remains stagnant for at least one hour

after rain, 9) produce that were commonly eaten raw, and 10) street food that was sold on the

street and commonly consumed by community members including children (S1 Table). All

drinking water samples were collected directly from the source and not from household stor-

age containers. We considered these to be priority environmental samples based on: 1) self-

reported behavior about contact and ingestion from people in the study neighborhoods, 2)

likelihood of contamination, as suggested by previous research in Bangladesh [12,14,15,17,18,

27–30], 3) recommendations from the stakeholders meeting, and 4) information from the

KIIs.

Sampling site selection

Before sample collection, the fieldworkers conducted a transect walk within each neighbor-

hood and noted possible sampling sites for each type of sample in all the neighborhoods. In

brief, for latrine swabs, fieldworkers purposively selected 10 shared/public latrines within each

neighborhood that met the inclusion criteria. If there were multiple latrines in a latrine block,

field workers selected the latrine that was reported to be the most frequently used. The field-

workers collected latrine swabs from the most frequently touched surfaces (i.e., 100 cm2 on flat

surfaces and door lock/handles) of the selected latrines. Fieldworkers collected 10 soil samples

in each neighborhood where children usually play. They also collected information on the type

of soil (soil/sand/mixed), distance between the closest latrine and the sample site, and whether

there were visible feces around the sampling area. For municipal and non-municipal drinking

and bathing water, the fieldworkers first purposively selected 10 shared water points of each

sample type in each neighborhood. Then, they recorded the source of the supplied water, type

of connection (legal/illegal) and secondary extraction source (shallow tubewell, deep tubewell,

public tap/standpipe, or piped water into the compound). Fieldworkers also measured the tur-

bidity (LaMotte Model 2020i, LaMotte Company, Chestertown, MD) and/or free chlorine

residual (LaMotte Model 1200, LaMotte Company, Chestertown, Maryland) of the water and

recorded the values using a mobile device. For drain water samples, fieldworkers explored all

open drains within the neighborhood during the transect walk and purposively selected 10

open drains where children play or where people came in to contact with the drain water while

walking. Floodwater samples were collected during the early monsoon (from June 1 to June

17, 2017). The fieldworkers collected the stagnant water that remained for at least one hour

after rainingfrom the street and/or courtyard where children play or people came in to contact

with the floodwater. Surface water samples were collected from the rivers, ponds, ditches, and/

or lakes within the neighborhoods where children often swim or play or people wash utensils/

clothing. During the transect walk, the fieldworkers explored all surface water sources in each

neighborhood and purposively selected 10 sources geographically distant from each other. If

the surface water source was small (pond/ditch), then the fieldworkers collected a single water

sample, and if the water source was large (lake/river), the fieldworkers collected multiple sam-

ples from different points of the same source. We collected prepared street foods from street

food vendors and/or from the street food shops depending on the availability during each day

of sample collection. Food items sold on the streets and commonly eaten by the children and

adults living in the community were collected. For this study, we collected Fuska (a round

puffed and fried pastry with a hole on the top to fill with spiced sauce), chotpoti (popular hot

SaniPath fecal exposure pathways urban Dhaka, Bangladesh

PLOS ONE | https://doi.org/10.1371/journal.pone.0221193 December 16, 2019 4 / 21

and sour snacks made of potatoes, chickpeas, onions, and chilies mixed with tamarind sauce),

and jhalmuri (mixture of puffed rice and a variety of spices including peanuts, mustard oil,

chili, onion, tomato, fresh ginger, salt, and lemon juice) (S2 Table) [15,31]. For produce, the

fieldworkers visited the local produce market in each neighborhood and sampled fresh pro-

duce that people commonly consumed raw or uncooked such as salad or garnish. Salads are

typically prepared with bare hands and consist of raw vegetables like tomatoes, cucumbers,

carrots, lettuce, coriander, onion, and green chili [32]. For this study, we collected samples of

tomatoes, cucumbers, and coriander leaves, which are common salad ingredients found in

Dhaka food markets.

Environmental sample collection

We standardized SaniPath protocols that have been applied in exposure assessments in several

other cities [33], to collect all environmental samples except for street food which was not

assessed during the previous SaniPath Tool assessments. After obtaining consent, the fieldwor-

kers requested the street food vendors to prepare a single serving as he/she usually prepares it.

The fieldworker held a 500 mL Whirl-Pak bag (Nasco, FortAtkinson, WI) with the mouth

open, and the vendor poured/placed the food into the bag. Only three municipal water samples

had detectable free residual chlorine (>0.20 mg/L), so we did not use the tablet containing 10

mg of active sodium thiosulfate to neutralize chlorine at the time of sample collection.

After each sample was collected, fieldworkers sealed the bag, noted the time of sample col-

lection, and immediately placed it into a cold box that was maintained at< 10˚C with ice

packs. Then, they used a mobile phone to record the Global Positioning System (GPS) coordi-

nates of the sampling site and take at least two photographs of the sample and/or sampling

site.

Laboratory sample processing

A laboratory supervisor received the environmental samples within 4 hours of collection and

analyzed the samples for E. coli using the IDEXX- Quanti-tray1 2000 technique with Colilert-

24 media (IDEXX Laboratories, Westbrook, Seattle, WA) [34] to quantify the most probable

number (MPN) of E. coli per unit of sample. E. coli is commonly used as an indicator of fecal

contamination in water, food, and environmental samples [13,35,36]. We chose to use E. colito allow for comparison with other studies.

Enumeration of E. coliAll environmental samples were processed on the same day, typically within 6 hours of collec-

tion, using the IDEXX Quanti-Tray 2000 system and Colilert reagent (IDEXX Laboratories,

Maine, USA). Initially, different dilutions of samples were pre-tested to determine the ideal

dilution factor to minimize samples with undetectable E. coli or E. coli concentrations exceed-

ing the Quanti-Tray upper detection limit. Due to the wide range of the sampling sites (high

income vs. low-income vs. floating), at least two dilutions per sample were analyzed to opti-

mize detection of positive E. coli wells within the Quanti-Tray detectable range of>1 to< =

2419.6 MPN per tray.

To assess the E. coli concentration in the environmental samples, the laboratory assistants

followed the SaniPath protocols for sample processing and analyses [33], except for street

foods. In brief, all drinking water (both municipal and non-municipal) and bathing water sam-

ples were analyzed without dilution and with a 1:10 dilution in distilled water. Surface water

and floodwater samples were diluted 1:102, 1:103 and 1:104 in distilled water; drain water was

diluted 1:105 and 1:106 in distilled water [37]. Street food samples were processed according to

SaniPath fecal exposure pathways urban Dhaka, Bangladesh

PLOS ONE | https://doi.org/10.1371/journal.pone.0221193 December 16, 2019 5 / 21

the protocol for food processing from the WASH Benefits Study and World Bank Add-on

Study in Bangladesh [38]. A 10g aliquot of street food was homogenized with 90 mL of distilled

water for one minute using a sterile BagMixer bag and BagMixer1 400 CC1 (Interscience

Laboratory Inc., Woburn, MA) at speed 4 with gap at-3 mm. We used MPN numbers from

1:10 and 1:102 dilutions to calculate the concentration of E. coli in the samples.For soil samples,

a 10g aliquot of soil was mixed with 20 mL Phosphate Buffered Saline with 0.04% of Tween 80

(PBST) [39] and homogenized with a shaker (Labnet, VX100 Vortex, MO BIO Laboratories,

Inc., MD) for 30 minutes. The homogenized samples were allowed to settle for another 30

minutes, and then diluted 1:102, 1:103 and 1:104in distilled water. Latrine swabs were

immersed in 14 mL of PBST and homogenized with a vortex (Labnet, VX100 Vortex, MO BIO

Laboratories, Inc., MD, USA) for 30 seconds, followed by incubation at room temperature for

5 minutes and mixed again for another 30 seconds. After mixing, the swab eluates were poured

into a 15 mL sterile conical tube and diluted 1:10 and 1:103indistilled water. For produce sam-

ples, whole pieces/amount of produce was rinsed with 500 mL PBST in a 2 L Whirl Pak stand-

up bag and were allowed to settle for 15 minutes. After mixing well, the laboratory assistant

removed the produce from the Whirl-Pak bag, weighed, and recorded the weight on a labora-

tory form. Then, the 10 mL eluates from the rinse sample were collected and diluted1:10, 1:102

and 1:103in distilled water.

One field blank of distilled water was collected and processed each day. The laboratory tech-

nician filled one 100 mL Whirl-Pak bag with distilled water in the study community as a mea-

sure of the staff’s sterile technique. This blank was then tested in the laboratory for E. coli, and if

the field blank showed any growth, we considered that contamination had occurred during

sample collection and reinforced aseptic precautions for subsequent sample collection. One lab-

oratory blank per laboratory assistant per day, one positive control (drain water), and one nega-

tive control (distilled water) per batch of Colilert per laboratory assistant per day were

processed for quality control. Only 2% (1/48) of the tested blanks, 4% (2/48) negative control

had positive growth, and all (N = 11) positive controls had E. coli growth. Finally, 100 mL envi-

ronmental samples were processed and sealed in a Quanti-Tray and incubated at 37˚C for 24

hours. The MPN of E. coli was determined by counting the number of fluorescing wells and cal-

culating according to the manufacturer instructions. All water samples were reported as MPN

of E. coli/100 mL, latrine swabs were reported as MPN of E. coli/swab, produce were reported as

MPN of E. coli/single serving, and street food samples were reported as MPN of E. coli/gram.

The weight (median, ranges) of produce and street food samples are provided in S4 Table.

Qualitative data analysis

The fieldworker who recorded Key Informant Interviews (KIIs) transcribed them in Bengali

so that thematic content analysis could be performed [40]. The investigator manually coded

the transcripts in an Excel spreadsheet according to the research objectives. After coding, the

investigator categorized the data under different themes and matched these themes to factors

associated with selection of environmental samples in each community.

Quantitative data analysis

We substituted the value of 0.5 MPN for samples below the detection limit and 2419.6 MPN

for samples above the detection limit, and calculated the E. coli concentration with corre-

sponding dilution factors (S2 Table). When the E. coli counts of all three dilutions were<1

MPN, we used the lowest diluted sample to estimate the concentration. When the E. colicounts of all three dilutions were>2419.6 MPN, we used the highest dilution to estimate the

concentration, and if at least one E. coli count was within the detectable limit (from 1 to 2419.6

SaniPath fecal exposure pathways urban Dhaka, Bangladesh

PLOS ONE | https://doi.org/10.1371/journal.pone.0221193 December 16, 2019 6 / 21

MPN) we calculated the average concentration of E. coli ignoring the censored (out of detect-

able limit) E. coli counts. If more than one dilution was positive and within the detectable

range (from 1 to 2419.6 MPN), we calculated the average concentration of E. coli from multiple

detectable trays. E. coli concentrations were log10 transformed, and summarized by sample

type and neighborhood. We compared E. coli contamination between the low-income, high-

income and floating neighborhoods (S3 Table), and between the north and south parts of the

city (DNCC and DSCC) using generalized linear regression models. We also examined differ-

ences in the level of contamination between neighborhoods graphically using an error bar

graph produced by R (version 3.4.1). We have conducted an analysis comparing the E. coliconcentrations in soil samples collected near visible feces within three meters to samples col-

lected in areas without visible feces using t-test and Wilcoxon rank-sum test. All statistical

analyses were conducted using STATA-13.

Ethics

The study protocol was reviewed and approved by the human subject research committees at

the International Centre for Diarrhoeal Disease Research, Bangladesh (icddr,b) [PR-17027]

and at Emory University [IRB00051584]. We obtained written informed consents from all

respondents during surveys and sample collection. The data were also analyzed anonymously

at both icddr,b and Emory University.

Results

Key informant interviews (KII)

City officials, and/or community leaders reported that shared latrines were the most common

type of latrine used by all communities. Key informants reported that open, rather than closed,

drains were common in all neighborhoods except Dhanmondi. The municipal water supply was

reported as the most common source of bathing water (direct municipal water 73%, tube well/

borehole 25%, dug well 2% and municipal water stored in a container 25%) and drinking water

throughout the neighborhoods except for Kamalapur and Uttarkhan. In Kamalapur, most people

used water from deep bore wells and commercially available 20L jar water, and in Uttarkhan, pri-

vate submersible pumpsconnected to a deep borehole were the main source of drinking water.

Commercially available jar waterwas also reported as the most commonly used drinking water in

all neighborhoods except for Uttarkhan. Almost all city officials, and/or community leaders

reported that fuska, chotpoti and jhalmuri were commonly eaten street foods and that cucumbers,

tomatoes, and coriander were commonly eaten raw vegetables in all neighborhoods (S4 Table).

Magnitude of E. coli contamination in environmental samples

Among environmental samples, almost all drain water (98%) and street food (93%) samples,

nearly 80% of fresh produce, surface water, soil and floodwater samples, and more than 50% of

municipal drinking water, non-municipal drinking water and bathing water samples were

contaminated with E. coli (Fig 1). Fifty seven percent of soil samples were collected where feces

were visible within 3 meters of sampling site and the concentrations of E. coli were not signifi-

cantly associated with presence of feces within 3 meters of sampling area. Characteristics of

individual samples are described in supplemental S5 Table (S5 Table).

Among the 10 neighborhoods, Hazaribagh had the greatest concentration of E. coli [mean

(SD)] in five categories of samples, including shared latrine swabs [0.64 log10 MPN/swab (0.99)],

municipal drinking water [3.20 log10 MPN/100 mL (0.84)], non-municipal drinking water [1.67

log10 MPN/100 mL (1.21)], surface water [7.38 log10 MPN/100 mL (0.00] and floodwater [5.47

SaniPath fecal exposure pathways urban Dhaka, Bangladesh

PLOS ONE | https://doi.org/10.1371/journal.pone.0221193 December 16, 2019 7 / 21

log10 MPN/100 mL (0.89)]). E. coli concentrations in drain water [7.61 log10 MPN/100 mL

(0.65)] from Shampur, street food [2.58 log10 MPN/gram (1.36)] from Dhanmondi, soil [3.19

log10 MPN/gram (1.49)] from Badda, produce [3.52 log10 MPN/serving (1.45)] from Kamalapur,

and bathing water [2.57 log10 MPN/100 mL (1.57)] from Shampur were higher than other neigh-

borhoods. The concentrations of E. coli in shared latrine swabs, bathing and municipal water,

and street food from Gulshan, and soil, non-municipal drinking water, surface water and flood-

water samples from Uttarkhan were the lowest compared to rest of the neighborhoods (Table 1).

Although overall concentration of E. coli in most of the sample types were similar between

DNCC and DSCC, E. coli concentrations were significantly higher in bathing water [log10

mean difference DNCC minus DSCC = -0.91 log10 MPN/100 mL (95% CI: -1.42, -0.41] and

municipal drinking water [log10 mean difference DNCC minus DSCC = -1.43 log10 MPN/

100mL (95% CI: -1.98, -0.89)] from DSCC compared to DNCC (Table 1 and Fig 1).

Overall, the municipal drinking water was more contaminated compared to non-municipal

water (mean difference: non-municipal minus municipal water = -0.73 log10 MPN/100 mL,

95% CI: -1.08, -0.37). Although, the E. coli concentrations were similar between the municipal

and non-municipal water in DNCC, the E. coli concentration was significantly higher in

municipal water in DSCC (mean difference: non-municipal drinking water minus municipal

drinking water = -1.30 log10 MPN/100 mL, 95% CI: -1.81, -0.79) (Table 2). As expected, sam-

ples of municipal water generally had lower concentrations of E. coli than bathing water, flood-

water, surface water and drain water.

Comparison of E. coli concentration across high-income, low-income, and

floating neighborhoods

Low-income vs. high-income neighborhoods. There was no significant difference in E.

coli contamination for latrine swabs, drain, municipal drinking water, produce and street

Fig 1. Percentage of E. coli positive environmental samples [(N = 704, DSCC = 378, DNCC = 326)] from10 study

neighborhoods in Dhaka city, 2017 ‡P<0.05: Significant (percentile) differences between Dhaka North City

Corporation (DNCC) vs. Dhaka South City Corporation (DSCC).

https://doi.org/10.1371/journal.pone.0221193.g001

SaniPath fecal exposure pathways urban Dhaka, Bangladesh

PLOS ONE | https://doi.org/10.1371/journal.pone.0221193 December 16, 2019 8 / 21

foods between low-income and high-income neighborhoods (Fig 2). The remaining five sam-

ple types had significantly higher E. coli concentrations in low-income neighborhoods [soil

(mean difference: low-income minus high-income = 0.91 log10 MPN/gram, 95% CI: 0.39,

1.43), bathing water (mean difference: low-income minus high-income = 0.98 log10 MPN/100

mL, 95% CI:0.41, 1.54), non-municipal water (mean difference:low-income minus high-

income = 0.64 log10 MPN/100 mL, 95% CI, 0.24, 1.04), surface water (mean differences from

low-income minus high-income = 1.92 log10 MPN/100 mL, 95% CI: 1.44, 2.40) and floodwater

Table 1. Mean log10 E. coli concentration in environmental samples from 10 Dhaka city neighborhoods, 2017.

Environmental Samples, Mean log10 MPN E. coli concentration (SD)

Neighborhoods Latrine

swabbSoild Drain

wateraBathing

wateraMunicipal

drinking wateraNon- municipal

wateraSurface

wateraProducec Street

fooddFlood

watera

N = 10 Samples/neighborhood

Floating communities

Gabtoli (N) 0.04 (0.63) 3.11

(1.02)

6.69

(0.97)

0.33

(0.87)

0.12

(0.64)

-0.15

(0.21)

5.77

(1.32)

3.37 (1.60) 1.79

(1.55)

4.21

(1.64)

Kamalapur (S) 0.79 (1.02) 3.06

(0.53)

6.23

(1.23)

1.62‡

(1.24)

1.16

(1.59)

0.67§ (0.84) 5.02

(1.28)

3.52

(1.45)

2.04

(1.0)

4.81

(0.74)

Unstructured slums

Kalshi (N) 0.05 (0.64) 2.26

(1.49)

6.64

(0.78)

2.12

(0.89)

0.16

(0.57)

0.69

(1.27)

6.05

(0.74)

2.57 (1.23) 1.44

(1.27)

4.57

(0.57)

Shampur (S) -0.06

(0.29)

2.00

(1.09)

7.61§

(0.65)

2.57

(1.57)

2.71|| (1.55) -0.19‡ (0.34) 5.86

(0.92)

3.55 (1.59) 0.76

(0.86)

4.52

(1.15)

Structured slums

Badda (N) 0.26 (0.65) 3.19

(1.49)

6.76

(1.10)

0.71

(0.92)

0.11

(0.72)

1.09

(0.93)

5.55

(0.92)

3.46 (1.25) 2.26

(1.02)

4.88

(0.39)

Hazaribag (S) 0.64

(0.99)

2.82

(1.01)

7.43

(0.41)

2.28‡ (1.06) 3.20||

(0.84)

1.67

(1.21)

7.38||

(0.01)

2.85 (1.06) 2.02

(0.66)

5.47

(0.89)

Non-slum with poor WASH

Uttarkhan (N) 0.36 (0.96) 1.08

(1.36)

7.00

(0.62)

1.07

(1.21)

1.96

(1.62)

-0.12

(0.25)

4.51

(1.29)

2.87 (1.29) 1.79

(1.21)

4.02

(0.68)

Motijheel (S) 0.12 (0.76) 2.29‡

(0.60)

6.68

(1.00)

1.62

(1.10)

1.57

(1.0.2)

0.32

(0.69)

3.43‡

(0.71)

3.14‡

(1.18)

1.92

(1.05)

4.81‡

(0.75)

Non-slum with improved WASH

Gulshan (N) -0.15

(0.01)

1.57

(1.30)

7.01

(0.49)

0.18

(0.87)

-0.47

(0.71)

0.01

(0.48)

5.07

(0.64)

3.23 (1.47) 1.34

(1.05)

4.62

(0.38)

Dhanmondi (S) -0.05

(0.26)

1.60

(1.02)

7.15

(1.14)

0.89

(1.45)

0.82

(1.33)

0.50

(0.72)

4.16§

(0.53)

3.36 (1.66) 2.58‡

(1.36)

4.07

(1.56)

Mean log10 E. coli in 10 study

neighborhoods

0.20 (0.73) 2.29

(1.29)

6.91

(0.92)

1.34

(1.34)

1.17

(1.55)

0.45

(0.94)

5.28

(1.37)

3.19 (1.36) 1.79

(1.18)

4.60

(1.02)

Mean difference DNCC

minus DSCC (95% CI)

-0.17

(-0.47,

0.11)

0.09

(-0.60,

0.42)

-0.20

(-0.57,

0.16)

-0.91

(-1.42,

-0.41)||

-1.43

(-1.98,

-0.89)||

-0.29

(-0.66, 0.08)

0.22

(-0.32,

0.77)

-0.18

(-0.72,

0.36)

-0.14

(-0.61,

0.33)

-0.27

(-0.68,

0.13)

N = Dhaka North City Corporation (DNCC), S = Dhaka South City corporation (DSCC)

‡Level of significance between neighborhoods at P<0.05

§Level of significance at P<0.01

||Level of significance at P<0.001, Bold digits: Highest mean log10 E. coli concentration in specific type of environmental sampleaAll water samples including drains were reported as MPN of E.coli/100 mLbLatrine swabs were reported as MPN of E. coli/swabcProduce were reported as MPN of E. coli/single servingdStreet food and soil samples were reported as MPN of E. coli/gram.

https://doi.org/10.1371/journal.pone.0221193.t001

SaniPath fecal exposure pathways urban Dhaka, Bangladesh

PLOS ONE | https://doi.org/10.1371/journal.pone.0221193 December 16, 2019 9 / 21

Table 2. Differences between E. coli concentrations per 100 mL in samples from municipal drinking water and other types of water in 10 neighborhoods in Dhaka

city, 2017.

Mean log10 MPN E. coli concentration differences from different water samplesvs. municipal drinking water

(95% CI)

Neighborhoods �DSCC (N = 50) †DNCC (N = 50) All 10 study neighborhoods

N = 100)

Municipal drinking water‡§ 1.89‡ 0.46‡ 1.18‡

Drain water vs. municipal water 5.12

(4.61, 5.64)||6.36

(5.96, 6.75)||5.74

(5.39, 6.09)||

Bathing water vs. municipal water -0.09

(-0.67, 0.48)

0.42

(-0.33, 0.88)

0.16

(-0.23, 0.57)

Floodwater vs. municipal water 2.84

(2.31, 3.37)||4.00

(3.59, 4.10)||3.42

(3.06, 3.79)||

Non-municipal vs. municipal water -1.30

(-1.81, -0.79)||-0.15

(-0.56, 0.25)

-0.73

(-1.08, -0.37)||

Surface water vs. municipal water 3.28

(2.66, 3.89)||4.93

(4.48, 5.39)||4.10

(3.69, 4.51)||

� Dhaka South City Corporation (DSCC)†Dhaka North City Corporation (DNCC)‡Reference value: E. coli contamination in municipal drinking water§The first row shows mean log10 E. coli concentration and the other columns show the mean log10 difference compared to municipal drinking water||mean differences were significantly different between the two comparison groups

https://doi.org/10.1371/journal.pone.0221193.t002

Fig 2. Mean log10 E. coli concentrations in environmental samples from 10 study neighborhoods in Dhaka City, 2017. Unit of measurements

used: all water samples (per 100 mL), latrine swab (per swab), produce (per single serving), street food (per gram), and soil (per gram). The

horizontal dotted line signifies the mean E. coli concentration of all samples (N = 1000). �Level of significance at P<0.05 †Level of significance at

P<0.01 ‡Level of significance at P<0.001 §Unit of measurement: all water samples (per 100 mL), latrine (per swab), produce (per single serving),

street food (per gram), Soil (per gram).

https://doi.org/10.1371/journal.pone.0221193.g002

SaniPath fecal exposure pathways urban Dhaka, Bangladesh

PLOS ONE | https://doi.org/10.1371/journal.pone.0221193 December 16, 2019 10 / 21

(mean difference: low-income minus high-income = 0.48 log10 MPN/100 mL, 95% CI: 0.03,

0.92)] (Table 3 and Fig 2).

Low-income vs. floating neighborhoods. Although concentrations of E. coli between

low-income and floating neighborhoods were similar for latrine swabs, soil, produce, street

food and floodwater samples, the concentrations were significantly higher in samples from

low-income neighborhoods compared to floating neighborhoods for drain water (mean

difference: low-income minus floating neighborhoods = 0.65 log10 MPN/100 mL, 95%

CI: 0.15, 1.14), bathing water (mean difference: low-income minus floating neighbor-

hoods = 0.94 log10 MPN/100 mL, 95% CI: 0.25, 1.63), municipal water (mean difference:

low-income minus floating neighborhoods = 0.91 log10 MPN/100 mL, 95% CI: 0.07, 1.73),

non-municipal water (mean difference: low-income minus floating neighborhoods = 0.55

log10 MPN/100 mL, 95% CI: 0.06, 1.04) and surface water (mean difference: low-income

minus floating neighborhoods = 0.81 log10 MPN/100 mL, 95% CI: 0.22, 1.40) (Table 3 and

Fig 2).

Floating vs. high-income neighborhoods. We found similar E. coli concentrations

between floating and high-income neighborhoods for all environmental sample types except

for soil (mean difference floating minus high-income neighborhoods: 1.46 log10 MPN/gram,

95% CI: 0.82, 2.10) and surface water (mean difference: floating minus high-income neighbor-

hoods = 1.10 log10 MPN/100 mL, 95% CI: 0.52, 1.69) (Table 3 and Fig 2).

Table 3. Comparisons between mean log10 MPN E. coli concentrations in environmental samples from low-income, high-income, and floating neighborhoods in

Dhaka city, 2017.

Neighborhoods Latrine

surface swabbSoild Drain

wateraBathing

wateraMunicipal

drinking wateraNon-municipal

wateraSurface

wateraProducec Street

fooddFloodwater a

Mean log10 MPN E.coliconcentration (SD)

Low-income�

(N = 40)

0.22 (0.72) 2.55

(1.34)

7.11

(0.86)

1.92 (1.32) 1.55 (1.73) 0.81 (1.19) 6.21 (1.01) 3.10

(1.13)

1.62

(1.11)

4.86 (0.88)

High-income†

(N = 40)

0.07 (0.63) 1.63

(1.16)

6.95

(0.84)

0.94 (1.23) 1.07 (1.41) 0.18 (0.60) 4.29 (1.01) 3.15

(1.37)

1.91

(1.21)

4.38 (0.98)

Floating‡ (N = 20) 0.42 (0.91) 3.09

(0.79)

6.46

(1.11)

0.99 (1.24) 0.64 (1.30) 0.26 (0.73) 5.40 (1.32) 3.45

(1.49)

1.92

(1.28)

4.51 (1.28)

Mean log10 MPN E. coli concentration differences between neighborhoods (95% CI)

Low-income vs.

high-income

0.15

(-0.17, 0.48)

0.91

(0.39,

1.43)§

0.16

(-0.25,

0.56)

0.98

(0.41, 1.54)§0.48

(-0.20, 1.16)

0.64

(0.24, 1.04)§1.92

(1.44,2.40)§-0.04

(-0.65,

0.56)

-0.29

(-0.81,

0.24)

0.48

(0.03, 0.92)§

Low-income vs.

floating

-0.19

(-0.59, 0.20)

-0.54

(-1.18,

0.95)

0.65

(0.15,

1.14)§

0.94

(0.25, 1.63)§0.91

(0.07, 1.73)§0.55

(0.06, 1.04)§0.81

(0.22,0 .40)§-0.35

(-1.09,

0.40)

-0.29

(-0.94,

0.34)

0.35

(-0.20, 0.89)

Floating vs. high-

income

0.35

(-0.05, 0.75)

1.46

(0.82,

2.10)§

-0.49

(-0.98,

-0.00)

0.03

(-0.65,0.72)

-0.43

(-1.26, 0.40)

0.08

(- 0.41, 0.57)

1.10

(0.52, 1.69)§0.30

(-0.44,

1.04)

0.01

(-0.64,

0.65)

0.13

(-0.42, 0.67)

�Low-income neighborhoods: Kalshi, Shampur, Badda, and Hazaribagh†High-income neighborhoods: Gulshan, Dhanmondi, Motijhil, and Uttarkhan‡Floating neighborhoods: Gabtoli and Kamalapur neighborhoods§ mean differences were significantly different between the two comparison groupsaAll water samples including drains were reported as MPN of E.coli/100 mLbLatrine swabs were reported as MPN of E. coli/swabcProduce were reported as MPN of E. coli/single servingdStreet food samples were reported as MPN of E. coli/gram.

https://doi.org/10.1371/journal.pone.0221193.t003

SaniPath fecal exposure pathways urban Dhaka, Bangladesh

PLOS ONE | https://doi.org/10.1371/journal.pone.0221193 December 16, 2019 11 / 21

Discussion

Extensive E. coli contamination was detected in most of the environmental samples collected

throughout the 10 urban study neighborhoods, suggesting that all residential areas of Dhaka

may be prone to fecal contamination regardless of geographic location or socio-economic sta-

tus. This is consistent with the prediction of the fecal waste flows analysis for Dhaka [11,41]

that estimated that 98–99% of fecal waste in Dhaka is ultimately distributed within the urban

environment–including residential areas. Our results confirm high levels of fecal contamina-

tion in multiple compartments of the residential urban environment of Dhaka. Few studies

have attempted to comprehensively measure fecal contamination in urban Dhaka. Previous

studies have focused only on specific pathways, but they have also reported high occurrence of

fecal contamination in environmental samples. A recent study in a large wholesale produce

market and neighborhood retail markets in Dhaka found that 100% of carrot and red ama-

ranth rinses, 92% of eggplant rinses, and 46% of tomato rinses were contaminated with E. coli[12]. Street-vended foods in Dhaka markets [14] and near schools (60% jhalmuri, 29% chot-poti) [15] were also reported to be highly contaminated with fecal bacteria. The detection of

fecal indicator bacteria in these environmental samples suggests the potential presence of path-

ogenic organisms and the potential risk of enteric disease among Dhaka residents who are fre-

quently exposed to these contaminated environments, drink contaminated municipal water,

and/or consume raw or undercooked produce or street foods [42].

Unlike previous SaniPath deployments [20] that focused primarily on low-income neigh-

borhoods, the Dhaka SaniPath assessment compared environmental contamination in a range

of high-income, low-income, and floating communities. This diversity of neighborhoods

allowed examination of fecal contamination that may be due to localized sources, such as a

contaminated surface water body, vs. fecal contamination that moves through the city among

both poor and wealthy neighborhoods through vehicles such as contaminated produce,

municipal piped water, or open drains. Our results suggest that, despite socio-economic and

infrastructure differences between the study neighborhoods, the fecal contamination levels for

some sample types, like drain water, municipal drinking water, produce, and street food, were

similar across neighborhoods. The widespread fecal contamination in these urban neighbor-

hoods may be due to unsafe fecal sludge management and consequent movement and distri-

bution of fecal contamination in the urban environment (i.e., through flooding, poor drainage

systems, and/or unsafe dumping of sludge) [3]. Previous analyses of existing sanitation data

concluded that<1% of household fecal sludge in Dhaka was effectively managed, and the vast

majority of waste water and fecal sludge was not contained and was either leaking out of pipes

and latrines or deliberately discharged directly into the environment [41]. Our primary data

collection confirms the presence of fecal contamination in the range of residential environ-

ments that we studied.

Conversely, non-municipal drinking water, bathing water, surface water, and soil samples

had significantly higher E. coli concentrations in low-income neighborhoods compared to

high-income neighborhoods and suggests that the contamination in these pathways may be

due to local sources of fecal discharge. Low-income urban neighborhoods are located mainly

in lower elevations and in the periphery of the city (i.e., Hazaribagh) [43], where flooding

occurs almost every year [44]. The floodwater runs off into storm sewers and ultimately into

surface water, and during heavy rainfall, the contaminated water returns to the environment

and contaminates the soil [45]. Poor drainage systems, improper child feces disposal, and poor

fecal sludge management likely increase the fecal contamination of the soil in low-income

neighborhoods [46]. Lastly, unimproved housing infrastructure (i.e., dirt floor/walkway), poor

hydraulic and physical integrity of the water distribution network (leaky flexible pipes and

SaniPath fecal exposure pathways urban Dhaka, Bangladesh

PLOS ONE | https://doi.org/10.1371/journal.pone.0221193 December 16, 2019 12 / 21

illegal connections), unsafe water storage, high population density, and poorly designed and

constructed on-site household and community sanitation systems that do not adequately con-

tain fecal sludge may contribute to higherlocalized fecal contamination levels in soil and water

in low-income neighborhoods in Dhaka [47,48].

The overall municipal water quality results reported here are consistent with previous stud-

ies in Dhaka that reported high levels of fecal contamination in municipal drinking water

mostly in low-income communities [17,49]. A nationally-representative water quality assess-

ment estimated that 41% of all improved water sources sampled across Bangladesh were con-

taminated with E. coli [50,51]. Piped water systems, which are almost exclusive to urban areas

of Bangladesh, were among the most contaminated drinking water sources. That assessment

also reported that 55% of the water samples from municipal public taps and more than 80% of

the samples from water taps on premises in urban neighborhoods of Bangladesh had E. colicontamination [11]. Contamination can occur either in the distribution system due to frequent

pipe breaks and illegal connections, low or negative water pressure due to intermittent service,

and/or because of poor domestic water storage structures and maintenance [28,52,53].

Although fecal contamination was widespread throughout urban Dhaka, we found signifi-

cantly higher concentrations of E. coli in most of the samples from DSCC compared to DNCC.

There are several possible explanations for these differences. Firstly, DSCC is an older part of

the city with older infrastructure (i.e., pipes, drainage) and lanes (narrow roads) that largely

lack a drainage system. These lanes often become flooded during rainfall [54]. Additionally,

the households of DSCC are closely packed together with a leaky water distribution system

and older sanitation facilities. Furthermore, the population density is about three times greater

in DSCC (>124,000 persons per square kilometer) compared to DNCC (<35,000 persons per

square kilometer) [55], and this presents an additional challenge to ensure adequate WASH

services in DSCC with limited resources. Finally, the highly polluted Buriganga River passes

beside DSCC and is a major source of environmental contamination.

Our results show that the low-income communities in DSCC had significantly higher E.

coli concentrations in their municipal water supply compared to the low-income communities

of DNCC (Table 1). In DSCC, the majority of the municipal water is distributed through the

Saidabad surface water treatment plant (with or without additional water treatment at the

neighborhood level by booster chlorination (ie. In-line chlorine injectors) and from deep bore

wells attached with or without booster chlorinator, and in DNCC, water is exclusively supplied

by deep bore wells [26]. The high concentration of E. coli in DSCC municipal water supply

may be due to the water residence time in a water distribution system [56] with compromised

physical and hydraulic integrity that allows intrusion of contamination. Additionally, recent

research on water quality in low-income urban communities in Dhaka reported that most of

the municipal water sources do not have chlorine injectors, chlorine injectors are often broken

and/or that the water was inappropriately treated before distribution [17,49].

High concentrations of fecal contamination have frequently been reported on produce in

low- and middle-income countries, including Bangladesh [12] and elsewhere [20, 57–60].

Fresh produce can be a vehicle for fecal contamination to move across the city to both poor

neighborhoods and high-income households [20] and can pose a major health risk to urban

populations [61]. Limited data are available on disease burdens attributed to food contamina-

tion in low- and middle-income countries [62,63]. The CDC estimates that nearly half of all

food-borne illnesses in the United States [64] are caused by contaminated fresh produce and

that more than 30% of gastroenteritis cases in low- and middle-income countries are linked to

food borne transmission [65]. The causes of the fecal contamination detected on the produce

in this study are not known and may be due to poor agricultural practices by farmers (e.g. use

of wastewater for irrigation) and unhygienic conditions in the produce markets. Observational

SaniPath fecal exposure pathways urban Dhaka, Bangladesh

PLOS ONE | https://doi.org/10.1371/journal.pone.0221193 December 16, 2019 13 / 21

studies in rural Bangladesh identified that produce washing practices during salad preparation

(uncooked and mashed cucumber, tomato etc.) within low-income neighborhoods were inad-

equate, and salads were often contaminated due to poor hygiene practices [32,66].

Over 90% of street food samples in this study were contaminated with E. coli, and there was

no geographical variation in the level of contamination. This is a major public health concern,

and a number of studies have reported that people who patronize street food vendors suffer

from food-borne diseases like diarrhea, cholera, typhoid fever, and other enteric diseases

[42,67,68]. A number of studies in Bangladesh [6,15,67,69,70] and elsewhere [42,68,71,72] also

found high levels of microbial contamination in street-vended foods. These foods can be con-

taminated in different ways. According to a government report, 94% of street food vendors in

Dhaka reported that they used the municipal water supply to prepare food and did not take

any measures to treat the water. The report also found that nearly 58% of the vendors did not

cover their food while selling and most vendors did not wash their hands with soap while pre-

paring the food [6]. Additionally, most of the vendors (68%) were located on footpaths; 30% of

vending carts were placed near drains; and 18% were placed near sewerage. A number of stud-

ies have suggested that street food vending sites could serve as breeding points for rodents,

insects, and flies and could promote proliferation of microorganisms and increase the risk of

food contamination and disease transmission [73, 74].

Strengths and limitations

This study is the most comprehensive and systematic assessment of fecal contamination ever

conducted in urban Dhaka and included not only a wide range of neighborhoods but also

examined 10 different types of environmental samples for fecal indicator bacteria. While this

study provides valuable information on both the magnitude of fecal contamination in the envi-

ronment and how it is distributed in the city, it also has some important limitations. First,

although a large number of environmental samples were collected from three types of neigh-

borhoods with different socio-economic status in an attempt to represent a range of condi-

tions, it was not possible to cover the entire city. Therefore, our findings may not be

generalizable to all urban neighborhoods in Dhaka, in Bangladesh, or to other cities in South

Asia, such as those with dry climates or with better fecal sludge management and improved

WASH facilities [20]. Additionally, while the sample size was appropriate for the primary

study objective of conducting an exposure assessment, it may not be sufficient for detecting

modest differences between individual neighborhoods or between environmental pathways.

Future studies of environmental contamination should increase the sample size for pathways

that have large variation and/or cover larger or more diverse geographical regions [75].

In this study, we measured E. coli as fecal indicator bacteria but did not attempt to detect

specific enteric pathogens in the environment–some of which survive longer than E. coli and

are highly infectious even at low concentrations. We are not able to estimate the disease bur-

den associated with the levels of fecal contamination that were detected in these neighbor-

hoods. A recent study in Dhaka suggested that multi-drug resistant (MDR) E. coli were

widespread in the public water supply in Dhaka, which could be potentially hazardous for

human health [76]. Further, we did not distinguish if the E. coli we detected were from ani-

mals, humans or possible environmental sources [77]. Although animal density in Dhaka

neighborhoods is low, a recent study reported that a ruminant-associated bacterial target was

detected in 18% of hand rinse and 27% of floor samples in a study neighborhood in Dhaka

[13]. A review in 2017 also suggested that exposure to animal feces in urban environments

may be associated with enteric diseases, soil-transmitted helminths infections, environmental

enteric dysfunction, and growth faltering [78]. These findings suggest that effective

SaniPath fecal exposure pathways urban Dhaka, Bangladesh

PLOS ONE | https://doi.org/10.1371/journal.pone.0221193 December 16, 2019 14 / 21

community fecal management should account not only for human sources of contamination

but also for animal sources in urban environments.

Conclusions and recommendations

The results of this study indicate that there is widespread fecal contamination in the public

domain in Dhaka in both low-income and high-income neighborhoods. The poor drainage sys-

tem, poor sanitation facilities, frequent flooding and poorly managed municipal water supply of

Dhaka may contribute to this extensive fecal contamination [79]. The evidence from this study

can inform policies and interventions to protect public health in Dhaka and can also identify

important research needs. Intervention strategies should consider how the geographic, infrastruc-

ture, and economic differences across the city impact various fecal exposure pathways and their

implications for effectively reducing fecal contamination in urban neighborhoods of Dhaka.

The high prevalence of municipal drinking water contamination reported here emphasizes

the importance of adopting appropriate organizational arrangements for the routine mainte-

nance and improvement of drinking water systems in order to prevent contamination in the

municipal piped network and alerting water utility and municipal authorities to problems with

the system that need to be addressed [80]. Appropriate, affordable, and effective centralized,

community-level, and household-level water treatment and storage technologies need to be

developed along with increased awareness among landlords and compound managers about

the importance of safe water management practices in both the public and private domain

[80]. Future studies should examine the excessive water contamination detected in DSCC and

identify the specific factors that contribute to this problem.

Of special concern is the evidence that the food supply in the city (fresh produce and street-

vended food) has high contamination levels and poses a risk citywide. This risk may be less vis-

ible than poor WASH infrastructure and therefore less targeted for intervention. Policies and

regulations for safe street food are weak and poorly enforced in most low- and middle-income

countries [81] and even non-existent in some countries [82] including Bangladesh [6]. There-

fore, formulation of appropriate food hygiene policies and proper enforcement are essential to

reduce therisks associated with street food consumption [71,83]. Further studies of the causes

of food contamination at farms, markets, and street vendors are needed to understand the crit-

ical points in the food production chain where contamination is introduced and how to pre-

vent this contamination and mitigate risk through changes in agricultural practices and food

handling and hygiene.

Improving fecal sludge management, training on food hygiene and produce handling for

food vendors, and improving the microbial quality of municipal water should be explored as

strategies to prevent the introduction of fecal contamination into different environmental

pathways. Long-term integrated programs that include the provision of urban WASH services,

housing/infrastructure improvement, behavior change communication, appropriate technol-

ogy development (i.e., safely-managed sanitation systems, online automated centralized and

community-level water treatment systems), improved food safety practices, and good personal

hygiene) [84], could reduce fecal contamination and improve the overall WASH conditions in

urban neighborhoods. Future studies should explore behavior that brings people into contact

with the environment and assess exposure to environmental fecal contamination through mul-

tiple pathways and the associated risks in different neighborhoods of urban Dhaka. Applica-

tion of sound risk analyses and formulation of appropriate environmental protection policies

are necessary to provide a strong scientific basis for the host of risk management options that

Dhaka city authorities may need to explore in order to ensure public health and safety and

achieve Sustainable Development Goal 6 (safely managed water and sanitation) by 2030.

SaniPath fecal exposure pathways urban Dhaka, Bangladesh

PLOS ONE | https://doi.org/10.1371/journal.pone.0221193 December 16, 2019 15 / 21

Supporting information

S1 Table. Definitions of different environmental samples collected for SaniPath deploy-

ment, 2017.

(DOC)

S2 Table. Dilutions tested of different types of environmental samples for laboratory anal-

ysis. �For Example: 1:100 dilution mean 1 unit of sample added with 99 units of PBS.

(DOC)

S3 Table. Definitions of neighborhoods for SaniPath Dhaka deployment, 2017.

(DOC)

S4 Table. Summary of information from the key informant interviews (KII) about the

characteristics of the 10 study neighborhoods in Dhaka city, 2017. Jhalmuri: mixture of

puffed rice and a variety of spices, including peanuts, mustard oil, chili, onion, tomato, fresh

ginger, salt and/or lemon juice

Chotpoti: popular hot and sour snacks among the urban people in Bangladesh, which is made

of potatoes, chickpeas, onions, and chilies mixed with tamarind sauce

Fuska: a round, puffed and fried crisp; a hole is created on the top to add a spiced sauce filling

Halim: made of wheat, barley, meat (usually minced beef or mutton), lentils and spices, some-

times rice is also used.

Dal puri: an unleavened deep-fried bread along with lentils, onions and/or (rock) salt.

(DOC)

S5 Table. Summary of characteristics of the environmental samples in all 10 study neigh-

borhoods of urban Dhaka, 2017.

(DOC)

S1 Fig. Locations of 10 neighborhoods� in Dhaka city, 2017. �Numbers within the map rep-

resent different neighborhood codes.

(TIF)

S1 File. Data set.

(DTA)

Acknowledgments

We are grateful to James Michiel, Senior mHealth and Informatics Analyst for his support at

the beginning of the project. We also acknowledge the efforts of the Dhaka SaniPath field

team: The study design, data collection and data entry were conducted by the SaniPath project

team: Badal Howlader, Khan AliAfser, Rana Mia, Shamim Ahamed, Abdul Barek, Mohammad

Rafik, Emdadul Haque, Raju Ahmed, and Md. Arifin. icddr,b is also grateful to the Govern-

ments of Bangladesh, Canada, Sweden and the UK for providing core/unrestricted support.

Author Contributions

Conceptualization: Nuhu Amin, Mahbubur Rahman, Suraja Raj, Shahjahan Ali, Jamie Green,

Solaiman Doza, Mahbub-Ul Alam, Tarique Md. Nurul Huda, Sabrina Haque, Leanne Uni-

comb, George Joseph, Christine L. Moe.

Data curation: Nuhu Amin, Suraja Raj, Jamie Green, Shimul Das, Yuke Wang, Mohammad

Aminul Islam.

Formal analysis: Nuhu Amin, Suraja Raj, Solaiman Doza, Mohammad Aminul Islam.

SaniPath fecal exposure pathways urban Dhaka, Bangladesh

PLOS ONE | https://doi.org/10.1371/journal.pone.0221193 December 16, 2019 16 / 21

Funding acquisition: Nuhu Amin, Mahbubur Rahman, Suraja Raj, Jamie Green, Leanne Uni-

comb, George Joseph, Christine L. Moe.

Investigation: Nuhu Amin, Suraja Raj, Shahjahan Ali, Shimul Das, Mohammad Aminul

Islam, Sabrina Haque, Leanne Unicomb, George Joseph, Christine L. Moe.

Methodology: Nuhu Amin, Mahbubur Rahman, Suraja Raj, Shahjahan Ali, Shimul Das, Solai-

man Doza, Yuke Wang, Mohammad Aminul Islam, Tarique Md. Nurul Huda, Sabrina

Haque, Christine L. Moe.

Project administration: Nuhu Amin, Shahjahan Ali, Jamie Green, Christine L. Moe.

Resources: Nuhu Amin, Jamie Green, Solaiman Doza, Yuke Wang, Sabrina Haque, Leanne

Unicomb, Christine L. Moe.

Software: Nuhu Amin, Jamie Green, Momenul Haque Mondol, Yuke Wang.

Supervision: Nuhu Amin, Mahbubur Rahman, Suraja Raj, Shahjahan Ali, Shimul Das, Solai-

man Doza, Mohammad Aminul Islam, Tarique Md. Nurul Huda, Leanne Unicomb,

George Joseph, Christine L. Moe.

Validation: Nuhu Amin, Suraja Raj, Jamie Green, Shimul Das, Momenul Haque Mondol,

Yuke Wang, Mahbub-Ul Alam, Tarique Md. Nurul Huda, Leanne Unicomb, George

Joseph, Christine L. Moe.

Visualization: Nuhu Amin, Momenul Haque Mondol, Yuke Wang, Christine L. Moe.

Writing – original draft: Nuhu Amin, Christine L. Moe.

Writing – review & editing: Mahbubur Rahman, Suraja Raj, Shahjahan Ali, Jamie Green, Shi-

mul Das, Solaiman Doza, Momenul Haque Mondol, Yuke Wang, Mohammad Aminul

Islam, Mahbub-Ul Alam, Tarique Md. Nurul Huda, Sabrina Haque, Leanne Unicomb,

George Joseph, Christine L. Moe.

References1. WHO (2006) Preventing Disease Through Healthy Environments: Towards an estimate of the environ-

mental burden of disease. Available at http://www.who.int/quantifying_ehimpacts/publications/

preventingdisease.pdf and accessed on Novvember 5, 2017.

2. Cairncross S, Valdmanis V, (2006) Water Supply, Sanitation, and Hygiene Promotion. In: Jamison DT,

Breman JG, Measham AR, et al., editors. Disease Control Priorities in Developing Countries. 2nd edi-

tion. Washington (DC): The International Bank for Reconstruction and Development / The World Bank;

2006. Chapter 41. Available at https://www.ncbi.nlm.nih.gov/books/NBK11755/pdf/Bookshelf_

NBK11755.pdf and accessed on November 5, 2017.

3. WSP (2014) The Missing Link in Sanitation Service Delivery: A Review of Fecal Sludge Management in

12 Cities. Available at https://www.wsp.org/sites/wsp.org/files/publications/WSP-Fecal-Sludge-12-City-

Review-Research-Brief.pdf and Accessed on January 24, 2018.

4. UN (2014) World Urbanization Prospects: The 2014 Revision. Available at https://esa.un.org/unpd/

wup/publications/files/wup2014-highlights.pdf and accessed on October 12, 2018.

5. Alam M-U, Winch PJ, Saxton RE, Nizame FA, Yeasmin F, Normen G, et al. (2017) Behaviour change

intervention to improve shared toilet maintenance and cleanliness in urban slums of Dhaka: a cluster-

randomised controlled trial. Tropical Medicine & International Health 22: 1000–1011.

6. Faruque Q, Haque QF, Shekhar HU, Begum S (2010) Institutionalization of Healthy Street Food System

in Bangladesh: A Pilot Study with Three Wards of Dhaka City Corporation as a Model.

7. Alam GMJ (1997) Environmental pollution of Bangladesh–It’s effect and control. Available at http://

www.bangladeshenvironment.com/index.php/polution-s/294-environmental-pollution-of-bangladesh-it-

s-effect-and-control and accessed on January 4, 2017. buet.ac.bd.

SaniPath fecal exposure pathways urban Dhaka, Bangladesh

PLOS ONE | https://doi.org/10.1371/journal.pone.0221193 December 16, 2019 17 / 21

8. Yeasmin F, Luby SP, Saxton RE, Nizame FA, Alam M-U, Dutta NC, et al. (2017) Piloting a low-cost

hardware intervention to reduce improper disposal of solid waste in communal toilets in low-income set-

tlements in Dhaka, Bangladesh. BMC Public Health 17: 682. https://doi.org/10.1186/s12889-017-

4693-x PMID: 28851334

9. Gunawan A, Gunawan A, Strande L (2015) SFD Promotion Initiative. Khulna Bangladesh. Final Report.

Dubendorf: Eawag. Available at https://www.dora.lib4ri.ch/eawag/islandora/object/eawag:14828 and

accessed on November 30, 2018.

10. SNV (2014) A Baseline Study to Assess Faecal Sludge Management of Residential Premises in

Selected Southern Cities of Bangladesh. Available at http://www.snv.org/public/cms/sites/default/files/

explore/download/snv_-_baseline_study_to_assess_fsm_of_residential_premises.pdf and accessed

on November 30, 2018.

11. Joseph G, Haque S, Yoshida N, Yanez Pagans M, Sohag MKUB, Moqueet N, et al. (2018) Promising

progress: a diagnostic of water supply, sanitation, hygiene, and poverty in Bangladesh: Main report

(English). WASH Poverty Diagnostic. Washington, D.C.: World Bank Group. Available at https://

openknowledge.worldbank.org/handle/10986/29450 and accessed on October 25, 2018.

12. Harris AR, Islam MA, Unicomb L, Boehm AB, Luby S, Davis J, et al. (2018) Fecal Contamination on Pro-

duce from Wholesale and Retail Food Markets in Dhaka, Bangladesh. Am J Trop Med Hyg 98: 287–

294. https://doi.org/10.4269/ajtmh.17-0255 PMID: 29165214

13. Harris AR, Pickering AJ, Harris M, Doza S, Islam MS, Unicomb L, et al. (2016) Ruminants Contribute

Fecal Contamination to the Urban Household Environment in Dhaka, Bangladesh. Environ Sci Technol

50: 4642–4649. https://doi.org/10.1021/acs.est.5b06282 PMID: 27045990

14. Islam MA, Mondol AS, Azmi IJ, de Boer E, Beumer RR, Heuvelink AE, et al. (2010) Occurrence and

characterization of Shiga toxin-producing Escherichia coli in raw meat, raw milk, and street vended

juices in Bangladesh. Foodborne Pathog Dis 7: 1381–1385. https://doi.org/10.1089/fpd.2010.0569

PMID: 20704491

15. Al Mamun M, Rahman SMM, Turin TC (2013) Microbiological quality of selected street food items

vended by school-based street food vendors in Dhaka, Bangladesh. Int J Food Microbiol 166: 413–

418. https://doi.org/10.1016/j.ijfoodmicro.2013.08.007 PMID: 24029025

16. Albert J, Luoto J, Levine D (2010) End-user preferences for and performance of competing POU water

treatment technologies among the rural poor of Kenya. Environ Sci Technol 44: 4426–4432. https://doi.

org/10.1021/es1000566 PMID: 20446726

17. Amin N, Crider YS, Unicomb L, Das KK, Gope PS, Mahmud ZH, et al. (2016) Field trial of an automated

batch chlorinator system at shared water points in an urban community of Dhaka, Bangladesh. J Water

Sanit Hyg Dev 6: 32–41.

18. Islam MS, Brooks A, Kabir MS, Jahid IK, Islam MS, Goswami D, et al. (2007) Faecal contamination of

drinking water sources of Dhaka city during the 2004 flood in Bangladesh and use of disinfectants for

water treatment. J Appl Microbiol 103: 80–87. https://doi.org/10.1111/j.1365-2672.2006.03234.x PMID:

17584454

19. Arnold BF, Colford JM Jr., (2007) Treating water with chlorine at point-of-use to improve water quality

and reduce child diarrhea in developing countries: a systematic review and meta-analysis. Am J Trop

Med Hyg 76: 354–364. PMID: 17297049

20. Robb K, Null C, Teunis P, Yakubu H, Armah G, Moe CL. (2017) Assessment of Fecal Exposure Path-

ways in Low-Income Urban Neighborhoods in Accra, Ghana: Rationale, Design, Methods, and Key

Findings of the SaniPath Study. Am J Trop Med Hyg 97: 1020–1032. https://doi.org/10.4269/ajtmh.16-

0508 PMID: 28722599

21. Wang Y, Moe CL, Null C, Raj SJ, Baker KK, Robb KA, et al. (2017) Multipathway Quantitative Assess-

ment of Exposure to Fecal Contamination for Young Children in Low-Income Urban Environments in

Accra, Ghana: The SaniPath Analytical Approach. Am J Trop Med Hyg 97: 1009–1019. https://doi.org/

10.4269/ajtmh.16-0408 PMID: 29031283

22. DCC (2011) Local Government (City Corporation) (Amended) Act. Available at http://old.dncc.gov.bd/

local-government-city-corporation-amended-act/ and accessed on February 6, 2019.

23. Bangladesh Bureau of Statistics (2015) Census of Slum Areas and Floating Population 2014, Bangla-

desh Bureau of Statistics (BBS). Available at http://203.112.218.65:8008/WebTestApplication/userfiles/

Image/Slum/FloatingPopulation2014.pdf and accessed on June 26, 2019.

24. Bangladesh Bureau of Statistics (2015) Census of slum areas and floating population 2014. Available at

http://203.112.218.65:8008/WebTestApplication/userfiles/Image/Slum/FloatingPopulation2014.pdf and

accessed on October 25, 2019.

25. Bangladesh Bureau of Statistics (2015) Education household survey 2014. Available at http://bbs.

portal.gov.bd/sites/default/files/files/bbs.portal.gov.bd/page/7b7b171a_731a_4854_8e0a_

f8f7dede4a4a/Educatin%20Household%20Survey%202014.pdf and accessed on October 25, 2019.

SaniPath fecal exposure pathways urban Dhaka, Bangladesh

PLOS ONE | https://doi.org/10.1371/journal.pone.0221193 December 16, 2019 18 / 21

26. Ministry of Foreigh Affairs (2015) Concept note for “Design and Implementation of the Saidabad III

Water Treatment Plant in Dhaka, Bangladesh.

27. Pickering AJ, Ercumen A, Arnold BF, Kwong LH, Parvez SM, Alam M, et. al. (2018) et al. Fecal contami-

nation along multiple environmental pathways is associated with subsequent diarrhea among children

in rural Bangladesh. Environ Sci Technol 14: 7928–7936.

28. Mahbub KR, Nahar A, Ahmed MM, A. Chakraborty (2011) Quality Analysis of Dhaka WASA Drinking

Water: Detection and Biochemical Characterization of the Isolates. J Environ Sci & Natural Resources

4: 41–49.

29. Amin MR, Ahmed W, M. K (2008) The status of safe drinking water and sanitation in Batabaria, Comilla,

Bangladesh. Ann Trop Med Public Health 1: 47–51.

30. Amin N, Pickering AJ, Ram PK, Unicomb L, Najnin N, Homaira N, et al. (2014) Microbiological evalua-

tion of the efficacy of soapy water to clean hands: a randomized, non-inferiority field trial. Am J Trop

Med Hyg 91: 415–423. https://doi.org/10.4269/ajtmh.13-0475 PMID: 24914003

31. Al Mamun M, Rahman SMM, Turin TC (2013) Knowledge and Awareness of Children’s Food Safety

Among School-Based Street Food Vendors in Dhaka, Bangladesh. Foodborne Pathog Dis 10: 323–

330. https://doi.org/10.1089/fpd.2012.1283 PMID: 23458028

32. Nizame FA, Leontsini E, Luby SP, Nuruzzaman M, Parveen S, Winch PJ, et al. (2016) Hygiene Prac-

tices during Food Preparation in Rural Bangladesh: Opportunities to Improve the Impact of Handwash-

ing Interventions. Am J Trop Med Hyg 95: 288–297. https://doi.org/10.4269/ajtmh.15-0377 PMID:

27296388

33. Emory University (2014) SaniPath Rapid Assessment Tool Manual. Available at http://sanipath.org/wp-

content/uploads/2015/09/SaniPath-Rapid-Assessment-Tool-User-Guide.pdf. Accessed on October 11,

2017.

34. IDEXX (2002) IDEXX Summary: evaluation of colilert and enterolert defined substrate Methodology for

wastewater application. Available athttp://www.idexx.co.uk/pdf/en_gb/water/64194006K.pdf and

accessed on February 1, 2018.

35. Boehm AB, Wang D, Ercumen A, Shea M, Harris AR, Shanks OC, et al. (2016) Occurrence of Host-

Associated Fecal Markers on Child Hands, Household Soil, and Drinking Water in Rural Bangladeshi

Households. Environ. Sci. Technol. Lett 3: 393–398.

36. Pickering AJ, Ercumen A, Arnold BF, Kwong LH, Parvez SM, Alam M, et al. (2018) Fecal Indicator Bac-

teria along Multiple Environmental Transmission Pathways (Water, Hands, Food, Soil, Flies) and Sub-

sequent Child Diarrhea in Rural Bangladesh. Environ Sci Technol 52: 7928–7936. https://doi.org/10.

1021/acs.est.8b00928 PMID: 29902374

37. Ercumen A, Pickering AJ, Kwong LH, Arnold BF, Parvez SM, Alam M, et al. (2017) Animal Feces Con-

tribute to Domestic Fecal Contamination: Evidence from E. coli Measured in Water, Hands, Food, Flies,

and Soil in Bangladesh. Environ Sci Technol 51: 8725–8734. https://doi.org/10.1021/acs.est.7b01710

PMID: 28686435

38. Doza S, Jabeen Rahman M, Islam MA, Kwong LH, Unicomb L, Ercumen A, et al. (2018) Prevalence

and Association of Escherichia coli and Diarrheagenic Escherichia coli in Stored Foods for Young Chil-

dren and Flies Caught in the Same Households in Rural Bangladesh. Am J Trop Med Hyg 98: 1031–

1038. https://doi.org/10.4269/ajtmh.17-0408 PMID: 29436348

39. Liao CH, Shollenberger LM (2003) Survivability and long-term preservation of bacteria in water and in

phosphate-buffered saline. Lett Appl Microbiol 37: 45–50. https://doi.org/10.1046/j.1472-765x.2003.

01345.x PMID: 12803555

40. Rahman A, Akter F (2019) Reasons for formula feeding among rural Bangladeshi mothers: A qualitative

exploration. PLOS ONE 14: e0211761. https://doi.org/10.1371/journal.pone.0211761 PMID: 30807588

41. Ross I, Scott R, Xavier JR (2016) Fecal Sludge Management: Diagnostics for Service Delivery in Urban

Areas. Case study in Dhaka, Bangladesh. Water and sanitation program (WSP); Water and sanitation

program technical paper. Washington, D.C.: World Bank Group.

42. Rane S (2011) Street Vended Food in Developing World: Hazard Analyses. Indian J Microbiol 51: 100–

106. https://doi.org/10.1007/s12088-011-0154-x PMID: 22282636

43. Berger M (2014) The Unsustainable City. Sustainability 6: 365.

44. Miguez M, Verol A, de Sousa M, Rezende O (2015) Urban Floods in Lowlands—Levee Systems,

Unplanned Urban Growth and River Restoration Alternative: A Case Study in Brazil. Sustainability 7:

11068.

45. Bhatta B (2010) Causes and Consequences of Urban Growth and Sprawl. In: Analysis of Urban Growth

and Sprawl from Remote Sensing Data. Advances in Geographic Information Science. Springer, Ber-

lin, Heidelberg. Available at http://www.springer.com/978-3-642-05298-9 and accessed on December

17, 2018.

SaniPath fecal exposure pathways urban Dhaka, Bangladesh

PLOS ONE | https://doi.org/10.1371/journal.pone.0221193 December 16, 2019 19 / 21

46. Blom K (2015) Drainage systems, an occluded source of sanitation related outbreaks. Archives of public

health = Archives belges de sante publique 73: 8–8. https://doi.org/10.1186/s13690-014-0056-6 PMID:

25722855

47. Parvin M, Begum A (2018) Organic Solid Waste Management and the Urban Poor in Dhaka City. Int J

Waste Resour 8.

48. UN-Habiate (2010) Solid waste management in the world’s cities water and sanitation in the world’s cit-

ies. Available at https://www.eawag.ch/fileadmin/Domain1/Abteilungen/sandec/E-Learning/Moocs/

Solid_Waste/W5/Solid_Waste_Management_World_cities_2010.pdf and accessed on December 05,

2018.

49. Pickering AJ, Crider Y, Amin N, Bauza V, Unicomb L, Islam SA, et al. (2015) Differences in Field Effec-

tiveness and Adoption between a Novel Automated Chlorination System and Household Manual Chlori-

nation of Drinking Water in Dhaka, Bangladesh: A Randomized Controlled Trial. PLOS ONE 10:

e0118397. https://doi.org/10.1371/journal.pone.0118397 PMID: 25734448

50. The World Bank (2018) Bangladesh: Access to Clean Water Will Reduce Poverty Faster. Available at

https://www.worldbank.org/en/news/press-release/2018/10/11/bangladesh-access-to-clean-water-will-

reduce-poverty-faster and accessed on February 01, 2019.

51. Bangladesh Bureau of Statistics, UNICEF (2015) Bangladesh Multiple Indicator Cluster Survey (MICS)

2012–2013. New York: UNICEF.

52. Sharma M, Alipalo M (2017) The Dhaka Water Services Turnaround: Asian Development Bank report.

Available at http://dwasa.org.bd/wp-content/uploads/2018/01/The-Dhaka-WASA-Water-Service-

turnaround-by-Manoj-Sharma-and-Melissa-Alipalo.pdf and Accessed on February 4, 2019.

53. Ali M, Lopez AL, You YA, Kim YE, Sah B, Miskery B, et al. (2012) The global burden of cholera. Bulletin

of the World Health Organization 90: 209–218A. https://doi.org/10.2471/BLT.11.093427 PMID:

22461716

54. Kabir A, Parolin B (2012) Planning and development of Dhaka- A story of 400 years. Available at http://

www.usp.br/fau/iphs/abstractsAndPapersFiles/Sessions/09/KABIR_PAROLIN.PDF and accessed on

January 01, 2019.

55. Bangladesh Bureau of Statistics (2011) Population and housing census: Socio-economic and demo-

graphic report. Available at http://203.112.218.66/WebTestApplication/userfiles/Image/BBS/Socio_

Economic.pdf and accessed on February 6, 2019.

56. Tinker SC, Moe CL, Klein M, Flanders WD, Uber J, Amirtharajah A, et al. (2009) Drinking water resi-

dence time in distribution networks and emergency department visits for gastrointestinal illness in Metro

Atlanta, Georgia. J water health. 7: 332–43. https://doi.org/10.2166/wh.2009.022 PMID: 19240359

57. Ensink JHJ, Mahmood T, Dalsgaard A (2007) Wastewater-irrigated vegetables: market handling versus

irrigation water quality. Tropical Medicine & International Health 12: 2–7.

58. Amoah P, Drechsel P, Abaidoo RC, Ntow WJ (2006) Pesticide and Pathogen Contamination of Vegeta-

bles in Ghana’s Urban Markets. Arch. Environ. Contam. Toxicol. 50: 1–6. https://doi.org/10.1007/

s00244-004-0054-8 PMID: 16328619

59. Vital PG, Dimasuay KGB, Widmer KW, Rivera WL (2014) Microbiological Quality of Fresh Produce

from Open Air Markets and Supermarkets in the Philippines. The Scientific World Journal 2014: 7.

60. Heaton JC, Jones K (2008) Microbial contamination of fruit and vegetables and the behaviour of entero-

pathogens in the phyllosphere: a review. J Appl Microbiol 104: 613–626. https://doi.org/10.1111/j.1365-

2672.2007.03587.x PMID: 17927745

61. Mills F, Willetts J, Petterson S, Mitchell C, Norman G (2018) Faecal Pathogen Flows and Their Public

Health Risks in Urban Environments: A Proposed Approach to Inform Sanitation Planning. International

Journal of Environmental Research and Public Health 15: 181.

62. Chlebicz A, Śliżewska K (2018) Campylobacteriosis, Salmonellosis, Yersiniosis, and Listeriosis as Zoo-

notic Foodborne Diseases: A Review. Int. J. Environ. Res. Public Health 15: 863.

63. Galate L., S. B (2015) Campylobacter—A Foodborne Pathogen. Int J Sci Res 4: 1250–1259.

64. Centers for Disease Control and Prevention (2015) Surveillance for Foodborne Disease Outbreaks

United States 2015 Annual Report. Available at https://www.cdc.gov/foodsafety/pdfs/

2015FoodBorneOutbreaks_508.pdf and accessed on September 10, 2018.

65. Scallan E, Kirk M, Griffin PM (2013) Chapter 1: estimates of disease burden associated with contami-

nated food in the United States and globally. In: Foodborne Infections and Intoxications. 4th edition.

Amsterdam, The Netherlands: Elsevier Inc., 3–18.

66. Nizame FA, Unicomb L, Sanghvi T, Roy S, Nuruzzaman M, Ghosh PK, et al. (2013) Handwashing

before Food Preparation and Child Feeding: A Missed Opportunity for Hygiene Promotion. Am J Trop

Med Hyg 89: 1179–1185. https://doi.org/10.4269/ajtmh.13-0434 PMID: 24080638

SaniPath fecal exposure pathways urban Dhaka, Bangladesh

PLOS ONE | https://doi.org/10.1371/journal.pone.0221193 December 16, 2019 20 / 21

67. Islam S, Nasrin N, Rizwan F, Nahar L, Bhowmik A, Esha SA, et al. (2015) Microbial Contamination of

Street Vended Foods from a University Campus in Bangladesh. Southeast Asian J Trop Med Public

Health 46: 480–485. PMID: 26521522

68. Kharel N, Palni U, Tamang JP (2016) Microbiological assessment of ethnic street foods of the Himala-

yas. Journal of Ethnic Foods 3: 235–241.

69. Khairuzzaman M, Chowdhury FM, Zaman S, Al Mamun A, Bari ML (2014) Food Safety Challenges

towards Safe, Healthy, and Nutritious Street Foods in Bangladesh. International Journal of Food Sci-

ence 2014: 9.

70. Khan M, Haque M, Jhorna D, Begum M (2013) Contamination of street food by Salmonella in Chitta-

gong City. J. Food Sci. Technol. Nepal 8: 81–83.

71. Alimi BA (2016) Risk factors in street food practices in developing countries: A review. Food Science

and Human Wellness 5: 141–148.

72. Newell DG, Koopmans M, Verhoef L, Duizer E, Aidara-Kane A, Sprong H, et al. (2010) Food-borne dis-

eases—The challenges of 20years ago still persist while new ones continue to emerge. Int J Food

Microbiol 139: S3–S15. https://doi.org/10.1016/j.ijfoodmicro.2010.01.021 PMID: 20153070

73. Umoh VJ, Odoba MB. (1999) Safety and quality evaluation of street foods sold in Zaria, Nigeria. Food

Control. 10:9–14.

74. Mensah P, Yeboah-Manu D, Owusu-Darko K, Ablordey A. (2012) Street foods in Accra, Ghana: how

safe are they? Bulletin of the World Health Organization. 80: 546–54.

75. Navab-Daneshmand T, Friedrich MND, Gachter M, Montealegre MC, Mlambo LS, Nhiwatiwa T, et al.

(2018) Escherichia coli Contamination across Multiple Environmental Compartments (Soil, Hands,

Drinking Water, and Handwashing Water) in Urban Harare: Correlations and Risk Factors. Am J Trop

Med Hyg 98: 803–813. https://doi.org/10.4269/ajtmh.17-0521 PMID: 29363444

76. Talukdar PK, Rahman M, Rahman M, Nabi A, Islam Z, Hoque MM, et al. (2013) Antimicrobial Resis-

tance, Virulence Factors and Genetic Diversity of Escherichia coli Isolates from Household Water Sup-

ply in Dhaka, Bangladesh. PLOS ONE 8: e61090. https://doi.org/10.1371/journal.pone.0061090 PMID:

23573295

77. Ishii S, Sadowsky MJ. (2008) Escherichia coli in the Environment: Implications for Water Quality and

Human Health. Microbes and Environments. 23: 101–8. https://doi.org/10.1264/jsme2.23.101 PMID:

21558695

78. Penakalapati G, Swarthout J, Delahoy MJ, McAliley L, Wodnik B, Levy K, et al. (2017) Exposure to Ani-

mal Feces and Human Health: A Systematic Review and Proposed Research Priorities. Environ Sci

Technol 51: 11537–11552. https://doi.org/10.1021/acs.est.7b02811 PMID: 28926696

79. Mahmud I, Mbuya N (2016) Water, Sanitation, Hygiene, and Nutrition in Bangladesh: Can Building Toi-

lets Affect Children’s Growth?. World Bank Study;. Washington, DC: World Bank.©World Bank.

https://openknowledge.worldbank.org/handle/10986/22800 License: CC BY 3.0 IGO.

80. World Health Organization (2011) Guidelines forDrinking-Water Quality: 4th edition. Available at https://

apps.who.int/iris/bitstream/handle/10665/44584/9789241548151_eng.pdf?sequence=1 and accessed

on March 4, 2019.

81. Liu Z, Zhang G, Zhang X (2014) Urban street foods in Shijiazhuang city, China: Current status, safety

practices and risk mitigating strategies. Food Control 41: 212–218.

82. Samapundo S, Climat R, Xhaferi R, Devlieghere F (2015) Food safety knowledge, attitudes and prac-

tices of street food vendors and consumers in Port-au-Prince, Haiti. Food Control 50: 457–466.

83. Alimi BA, Workneh TS (2016) Consumer awareness and willingness to pay for safety of street foods in

developing countries: a review. International Journal of Consumer Studies 40: 242–248.

84. Jaitman L, Jose B (2013) Evaluation of slum upgrading programs: literature review and methodological

approaches. Available at https://pdfs.semanticscholar.org/d866/d29170a2a0c968ec07dfccdb6e782

bb3db15.pdf and accessed on January 1, 2019. Inter-American Development Bank

SaniPath fecal exposure pathways urban Dhaka, Bangladesh

PLOS ONE | https://doi.org/10.1371/journal.pone.0221193 December 16, 2019 21 / 21